1. technology and computing
2. software
3. databases

MULISA : A NEW STRATEGY FOR
DISCOVERY OF PROTEIN FUNCTIONAL
MOTIFS AND RESIDUES
Yi-Yuan Chiu1, Chih-Ta Lin1, Chien-Heng Lin1 and Jinn-Moon Yang1,2
1
Institute of Bioinformatics and Systems Biology,
National Chiao Tung University, Hsinchu, Taiwan.
2
[email protected]
[email protected]
Department of Biological Science and Technology,
National Chiao Tung University, Hsinchu, Taiwan.
[email protected]
ABSTRACT
To predict and identify details regarding function from protein sequences is an emergency task
since the growing number and diversity of protein sequence. Here, we develop a novel approach
for identifying conservation residues and motifs of ligand-binding proteins. In this method,
called MuLiSA (Multiple Ligand-bound Structure Alignment), we first superimpose the ligands
of ligand-binding proteins and then the residues of ligand-binding sites are naturally aligned.
We identify important residues and patterns based on the z-scores of the residue entropy and
residue-segment entropy. After identifying new pattern candidates, the profiles of patterns are
generated to predict the protein function from only protein sequences. We tested our approach
on ATP-binding proteins and HEM-binding proteins. The experiments show that MuLiSA can
identify the conservation residues and novel patterns which are really correlated with protein
functions of certain ligand-binding proteins. We found that our MuLiSA can identify
conservation patterns and is better than traditional alignments such as CE and CLUSTALW in
some ligand-binding proteins. We believe that our MuLiSA is useful to discover ligand-binding
specificity-determining residues and functional important patterns of proteins.
KEYWORDS
Multiple ligand-bound structure alignment, Functional motifs, Conserved patterns, ATP-binding
proteins, HEM-binding proteins.
1. INTRODUCTION
Human genome have been sequenced and led to a flood of sequence information. On the other
hand, recent developments in X-ray crystallography and NMR have made it faster in solving
protein structures. These data contains a lot of information that can be extracted by techniques
which were used to visualize the sequence conservation information. The residues most related to
the functions of a protein are often the most conserved (1). Many studies have demonstrated that
most protein domains of same protein families, such as PROSITE (2) and Pfam (3), share
conserved peptide patterns, called motifs, and some critical residues. The fundamental problems
in proteomics include both identifying and understanding the role of the essential sites that
determine that structure and proper function of the proteins. After solving these problems,
Jan Zizka et al. (Eds) : CCSEIT, MoWiN, IT, AIAP, ICBB - 2015
pp. 75–88, 2015. © CS & IT-CSCP 2015
DOI : 10.5121/csit.2015.51109
76
Computer Science & Information Technology (CS & IT)
researchers can apply this useful information as a clue to predict protein functions without protein
structure information.
Many groups have used the identification of conserved patterns as a method to predict protein
function. Some of these groups predict protein motifs using principle component analysis (4-7).
Other groups use structure alignment (8) or sequence alignment (9) as a method to identify
conservation sites. Evolutionary trace analysis was used to predict functional patterns in different
phylogenic trees and look for functional important residues (8, 10-13). However, these methods
always use protein structure or protein sequence information to predict protein conservation
patterns and may miss these conservation patterns because of the noises from other protein
structures which are far apart from ligand-binding site.
Figure 1. The workflow of analysis and identification of conservation patterns and residues in proteins by
MuLiSA. This flow starts from dataset preparation and clustering, followed by multiple ligand-bound
structure alignments (MuLiSA), tool evaluation and protein function prediction.
2. MATERIALS AND METHODS
Identification of conservation patterns and residues in proteins by multiple ligand-bound structure
alignments encompasses a variety of sequential computational phases, including dataset
preparation, dataset clustering, multiple ligand-bound structure alignments, post-alignment
analysis and entropy calculation, tool verify and protein function prediction (Figure 1). In dataset
preparation, we first select one kind of ligand-binding protein that we are interested and get
ligand-binding protein list from PDBsum (18) database. Because we need precise protein
structures to identified conservation residues and motifs, we only select protein structures
resolved by X-ray diffraction. Then we select ligand-binding domains using programs from
SCOP database (19). In data clustering, we generate all-against-all multiple ligand-bound
structure alignments of these selected ligand-binding domains and generate one structure
similarity matrix and one un-gapped sequence identity matrix for each kind of ligand-binding
proteins. Once we have these two matrixes, we select non-redundant protein domains, and
undergo protein domain clustering.
Computer Science & Information Technology (CS & IT)
77
In the main step of MuLiSA, first we choose the alignment center domains C of each domain
cluster based on structure similarity. Second, we undergo C centered multiple ligand-bound
structure alignment. After we generate the alignments, z-score calculation of position entropy can
help us to identify conservation residues of each domain cluster. For we believed that the
functional important motifs mostly composed of functional important residues, we identified
pattern candidates by conservation residues extension. Finally, we used SCOP (19) and PROSITE
(2) databases to verify our results; and then we generate profiles of pattern candidates and use
them to search for protein sequence with these patterns in SWISS-PROT database (20).
2.1. Preparation of ligand-binding proteins
We have applied MuLiSA to three kinds of ligand-binding proteins, which are ATP-binding
proteins, ADP-binding proteins, and HEM-binding proteins. The ligand-binding protein lists were
taken from PDBsum database (18). In order to get ligand-binding domains, first we need to get
ligand-binding protein structures. Protein structure three-dimensional information was
downloaded from Protein Data Bank (PDB) database(21) according to ligand-binding protein lists
getting from PDBsum database (18). The ligand-binding domains were chosen downloaded from
Structure Classification of Proteins (SCOP) database (19).
Ligand-binding domains were chosen with four criteria, they are as follows: 1) When one of
distances between atoms of residues of the domain and atoms of ligands is near than 5Å, we think
that this domain is a ligand-binding domain. 2) Because multiple ligand-bound structure
alignment first superimposed the ligands of aligned proteins, we only choose protein domains
which only bind with one ligand. 3) We only choose ligand-binding domains which the ligand
they bind is only bind by one protein domain. 4) We only choose one protein domain in one
protein structure. Because the SCOP domain files do not contain ligand information, after
choosing these domains we must add back ligand information from Protein Data Bank (PDB)
database (21) into these protein domain files. It must be mentioned that we only choose protein
domains solved by x-ray crystallography because we think that these structures are more
convincing.
2.2. Datasets for verification
To verify whether our alignment results is reasonable and can reflect protein function
information, we use the classification of Structural Classification of Proteins (SCOP) database
(19) as the benchmark of our structure similarity matrix for non-redundant domain clustering.
PROSITE patterns from PROSITE database (2) were also used to quality assessment and
refinement of multiple ligand-bound structure alignments. The protein sequences and annotations
were downloaded from SWISS-PROT database (20) and were used for profile verification and
protein function prediction.
2.3. Method
The main idea of this tool is that we try to align together conservation residues of proteins at
ligand-binding sites by ligand superimposition; and then identify conservation residues and
patterns by z-score of entropy calculation. Because we have to change the three-dimensional
coordinates of proteins along with superimposed ligands, we developed a structure superimpose
tool to deal with this problem. We developed this program MuLiSA from ICP algorithm(22), this
program can make proteins and ligands rotation and displacement on three-dimensional space.
After we get the superimposed protein structures, we regard two residues are aligned together
based on three order rules: 1) Rule 1: Cβ or Cα (Gly) atom of amino acid residues in 1Å; 2) Rule
2: Cβ or Cα (Gly) atom of same amino acid residues in 4Å; 3) Rule 3: Cβ or Cα (Gly) atom of
78
Computer Science & Information Technology (CS & IT)
same group amino acid residues in 4Å or Cβ or Cα (Gly) atom of different group amino acid
residues in 2Å.
The amino acid groups are defined as follows: 1) Basic amino acids: lysine, arginine, and
histidine; 2) Acidic amino acids: aspratate, glutamate, asparagine, and glutamine; 2) Aromatic
amino acids: phenylalanine, and tryptophan; 3) Aliphatic amino acids: glycine, alanine, valine,
leucine, isoleucine, and methionine; 4) Hydroxyl containing amino acids: serine, threonine, and
tyrosine; 5) Disulfide-bond forming amino acid: cysteine; 6) Cyclic amino acid: proline.
2.4. Sequence identity matrix and structure similarity matrix
If two protein domains have the similar function and have highly similar structures in ligandbinding sites, these two protein domain structures should fit well in three-dimensional space. We
introduced structure similarities in accordance with multiple ligand-bound structure alignments to
present this information. STab is the structure similarity of protein domain a and protein domain b.
La is the length (residue numbers) of protein domain a, Lb is the length (residue numbers) of
protein domain b, and L is the aligned residue number of protein domains a and b. STab is given as
L
T
. We also generate un-gapped sequence identity matrix between protein domains
=
S
ab
min{La , Lb}
for non-redundant protein domain selection based on only aligned residues of protein domains a
and bSEab is the un-gapped sequence identity of protein domain a and protein domain b. mt is the
number of identical aligned residues of protein domain a and protein domain b; mmt is the
number of non-identical aligned residues of protein domain a and protein domain b and
mt
E
.
S =
ab
mt + mmt
Redundant protein domains must be removed because the profiles generated from alignments
may be incredible. We regarded two protein domains are redundant protein domains when their
structure similarity and sequence identity are both above 0.8; therefore, we first cluster these
protein domains and only choose one with no mutation residues and with the smallest X-ray
diffraction resolution. In order to generate a convincing multiple alignments, we must choose an
alignment center domains C before we generate this alignments. In structure similarity matrixes,
the non-redundant protein domain of one cluster which has the highest structure similarity with
other protein domains than others was selected as the alignment center C of this cluster. This
protein domain was used to be the alignment center of multiple ligand-bound structure alignment.
2.5. Identification of conservation residues and pattern candidates
To identify these conservation residues, we used entropy (Sp), defined as
20
S
P
= −∑
i =1
f
pi
* ln
( f ),
pi
where i and fpi denote the i amino acid type, the probability of finding the amino acid type i at
position p. The entropy is 0 when this position is totally conserved. In order to estimate the
statistical significance of the position entropy, z-score was applied to identify relative
X p − µ , where Zp is the z-score value of position p, σis the standard
conservation positions:
=
th
Z
p
σ
deviation of all positions entropy,µis the average value of all positions entropy and Xp is the
entropy of position p. We identified a conservation position p when Zp > 2.5.
We generate alignment profiles of pattern candidates (discovered by our MuLiSA) and PROSITE
patterns from multiple ligand-bound structure alignments. PF = f i where 1 ≤ i ≤ 20 , where
pi
{ }
p
PFp is the profile of position p; fp is the probability of the i amino acid type at position p.We
i
th
Computer Science & Information Technology (CS & IT)
79
then used profiles to search for matched protein segments in protein sequences. The search
window size is the length of profiles and shifts one residue each time. Each protein sequence
should have N-(n-1) (N is the length of this sequence and n is the length of this pattern) profile
search scores, and we suppose the segment with the highest profile search score of this protein
sequence should be the pattern candidate that we are looking for. The scoring function is as
n
follows:
20
∑∑ PF
S=
p =1 i =1
pi
, where S is the profile search score, n is the length of a pattern, PFpi is the
n
profile value of amino acid type i at position p. The score is 1 when a segment perfectly matches
this profile.
Table 1. Statistics of proteins, domains and pattern candidates
Ligand
name
ATP
No. of
proteins
a
173
No. of
selected
domains
b
60
No. of nonredundant
domains c
Clusters
46
Protein kinases
catalytic subunit
(7)
Class I
aminoacyl-tRNA
synthetases (RS),
catalytic domain
(4)
d
CCP-like (13)
Heme
1145
860
131
Cytochrome
P450 (13)
Cytochrome b5
(5)
Monodomain
cytochrome c
(23)
Cytochrome c'
(4)
No. of
pattern
candidates
Selected
alignment
center C e
No. of
important
residues f
d1phk__
10
1
d1maua_
16
3
d1llp__
11
3
d1eupa_
12
3
d1cyo__
4
1
d1i54a_
3
1
d1i54a_h
3
1
g
Number of ligand-binding proteins in PDBsum database.
Number of ligand-binding domains selected by our program.
c
Number of selected non-redundant domains.
d
The domain clusters that according to structure similarity and SCOP database classification; the domain
names are based on SCOP database nomenclature. We only choose domain clusters with domain number
> 3 because the alignments are more statistical meaningful; and we only choose domain clusters with
PROSITE patterns because we need benchmarks to verify our results. The numbers in the parentheses are
the non-redundant domain numbers of each cluster.
e
The alignment center C domain of each cluster. The alignment center C chosen is important; because in
MuLiSA, the alignment center C highly affects the alignment results.
f
Number of conservation residues with z-score > 2.5.
g
Number of identified pattern candidates with length equal or longer than 5 residues.
h
We choose same alignment center C of domain clusters: monodomain cytochrome c and cytochrome c',
because same pattern candidates were identified in these clusters.
a
b
80
Computer Science & Information Technology (CS & IT)
3. RESULTS
In order to identify the wealth of information present in protein structures, we analyzed
conservation residues and patterns in multiple ligand-bound structure alignments. Most sequence
and structure alignment techniques are protein-based alignment; in other words, these techniques
analyze residue conservation only by comparing protein structure or protein sequence similarity.
At the present, we have applied MuLiSA to ATP- and HEM-binding proteins and identified
several conservation residues and pattern candidates. We have generated sequence profiles from
multiple alignments and used them to discover protein sequences which may have these profiles.
We also proved that MuLiSA is better than other tools in several cases and can discover
functional information when comparing with SCOP (19) and PROSITE database(2). Our major
intention was to extract protein structure information from ligand-binding proteins and apply this
information to protein function prediction. Table 1 shows some statistics about the dataset we
used in this study. Through getting ligand-binding protein lists, selecting ligand-binding domains,
domain clustering, non-redundant domains and alignment center C selection, we use MuLiSA and
z-score of entropy calculation to identified conservation residues and pattern candidates of each
cluster. These identified conservation residues may be functional important and we survey the
literature and it proves that some of these identified conservation residues are critical to ligandbinding or correlate with conformation stability. After pattern candidate identification, we
generate profiles of these pattern candidates and use these profiles predict protein functions.
3.1. ATP-binding proteins
ATP, adenosine triphosphate, is the major energy currency of the cell. It transfers energy from
chemical bonds to endergonic reactions of the cell. ATP powers most of the energy-consuming
activities of cells, such as muscle contraction, synthesis of polysaccharides, active transport of
ions and nerve impulse. Because of ATP is a so important compound and because of the large
number of experimental data, like ATP-binding protein structures and literatures, we choose
ATP-binding proteins as our first research target. We have generated structure similarity matrix
of non-redundant ATP-binding domains for functional-based domain clustering, and we also
identified conservation residues and pattern candidates. Finally, we used profiles of pattern
candidates to undergo protein function prediction.
3.1.1 Structure similarity matrix and alignment center selection
Figure 2 shows the structure similarity matrixes and SCOP classifications of 25 non-redundant
ATP-binding domains. When comparing with classifications of SCOP database (19), protein
domains with higher structure similarities are usually clustered together and they are always
belong to same SCOP families. As we all agree that SCOP database (19) is a convincing domain
structural and functional classification database, it tells us that the multiple ligand-bound
alignment and structure similarity calculation is reasonable and can reflect structural and
functional information.
In Figure 2A, the domains belong to the same SCOP families are with same colors. The bold
values means the structure similarity is larger than the average value of the row; in other words,
the domain in this row is much similar with these compared domains than others. In this matrix,
we find that most domains of same SCOP family usually have higher structure similarity with
each other (see the regions with red frame), it tell us that the multiple ligand-bound structure
alignment and structure similarity calculation is reasonable and can reflect structural and
functional information. Figure 2B shows the SCOP classification of protein domains.
Computer Science & Information Technology (CS & IT)
81
The alignment center C chosen is important; because in MuLiSA, the alignment center C highly
affects the alignment results. The alignment center C of one cluster is chosen when one domain
has higher structure similarity with other domains than others in this cluster. For example, the
first cluster in Figure 2A, “Class I aminoacyl-tRNA synthetases (RS), catalytic domain family”,
contains four domains: d1gtra2, d1h3ea1, d1maua_and d1n77a2. The structure similarity
summation of d1gtra2 with others is 0.3+0.39+0.34=1.03; d1h3ea1 is 0.3+0.44+0.36=1.1;
d1maua_is 0.39+0.44+0.36=1.19; and d1n77 is 0.34+0.36+0.36=1.06. Therefore, we choose
d1maua_ as the alignment center C of this cluster.
A
B
Figure 2 (A) Structure similarity matrix of 25 non-redundant ATP-binding domains; (B) SCOP
classification of 25 non-redundant ATP-binding domains. The domains belong to same SCOP families are
with same colors. The bold values means the structure similarity is larger than the average value of the row;
in other words, the domain in this row is much similar with these compared domains than others. In this
matrix, we find that most domains of same SCOP family usually have higher structure similarity with each
other (see the regions with red frame), it tells us that the multiple ligand-bound structure alignment and
structure similarity calculation is reasonable and can reflect structural and functional information. The
protein domains were classified according to SCOP classification hierarchy: class, fold, superfamily, and
family. The protein domains were named by SCOP database nomenclature.
3.1.2 Protein functional motifs and residues
Figure 3 shows the multiple ligand-bound structure alignment results and the identified
conservation residues in “Protein kinases, catalytic subunit family” of ATP-binding domains. The
identified conservation residues, aligned positions with z-score of entropy calculation > 2.5, are
82
Computer Science & Information Technology (CS & IT)
close to ATP in three-dimensional space. It implies that these conservation residues may play
important role in ATP-binding. In Figure 3B, the labeled residue numbers belong to protein
domain d1phk__, which is the selected alignment center C of this cluster; and the red framed
region means the PROSITE patterns. We observed that most identified conservation residues
were on these PROSITE pattern region, it tell us that identifying pattern candidates from
conservation residues extension may be a reasonable approach.
A
B
Figure 3. MuLiSA result and identified conservation residues in “Protein kinases, catalytic subunit family”
of ATP-binding domains. (A) Three-dimensional distributions of identified conservation residues and the
ligand superimposition. Yellow: d1phk__; blue: d1atpe_; green: d1qmza_; red: d1csn__; grey: d1hck__;
pink: d1gol__; light blue: d1h1wa_; (B) Multiple ligand-bound structure alignment result of “Protein
kinases, catalytic subunit family” domains. The identified conservation residues, aligned positions with zscore of entropy calculation > 2.5, are close to ATP in three-dimensional space. It implies that these
conservation residues may play important role in ATP-binding. The labeled residue numbers belong to
protein domain d1phk__, which is the selected alignment center C of this cluster; and the red framed region
means the PROSITE patterns. We observed that most identified conservation residues were on these
PROSITE pattern region, it tell us that identifying pattern candidates from conservation residues extension
may be a reasonable approach.
3.1.3 Protein functional pattern verifications
In order to verify the effectiveness of profiles generated from our alignments in protein function
prediction, we compare the performance in profile search between dataset 1, which contains
protein sequences with PROSITE pattern; and dataset 2, which contains protein sequences not
only with PROSITE pattern but also have “ATP-binding” annotations in SWISS-PROT database.
Dataset 1 contains protein sequences contain PROSITE pattern: aminoacyl-transfer RNA
synthetases class-I signature and dataset 2 contains protein sequences contain not only PROSITE
pattern: aminoacyl-transfer RNA synthetases class-I signature but also have “ATP-binding”
annotations in SWISS-PROT database. We observed that the area under curves of dataset 2 is
larger than the area under curves of dataset 1. Because the profile of pattern candidates were
generated from alignments of ATP-binding domains and the protein sequences in dataset 1 are
not all have “ATP-binding” annotations in “KW” of SWISS-PROT database, we suppose that the
Computer Science & Information Technology (CS & IT)
83
profile of pattern candidate is more convincing in ATP-binding proteins but not proteins only
with PROSITE patterns.
Table 2. Hit rate comparison of dataset difference in profile verification of ATP-binding proteins
Family
PROSITE patterns and
pattern candidates a
Dataset 1 c
No. of
sequence e
Protein kinases ATPbinding region signature
Protein kinases
catalytic subunit
Class I aminoacyltRNA synthetases
(RS), catalytic
domain
Serine/ Threonine protein
kinases active-site
signature.
Pattern candidate 1 b
Pattern candidate 2
Pattern candidate 3
Aminoacyl-transfer RNA
synthetases class-I
signature
Pattern candidate 1
Hit rate f
Dataset 2 d
No. of
sequence
85.15%
859
85.73%
89.18%
773
84.79%
64.19%
71.37%
1129
26.61%
20.18%
Hit rate
86.67%
86.76%
68.35%
75.43%
1056
50.42%
37.43%
PROSITE patterns and pattern candidates that we identified.
Pattern candidate 1 of “Protein kinases catalytic subunit family”.
c
Dataset 1: sequences only with PROSITE patterns
d
Dataset 2: sequences with PROSITE patterns and SWISS-PROT annotations
e
Number of sequences recorded which have PROSITE patterns in this cluster. For example, in “Protein
kinases catalytic subunit family”, number of sequences which have Protein kinases ATP-binding region
signature or Serine/ Threonine protein kinases active-site signature is 859; hence the number of sequences
of pattern candidates in this cluster is 859.
f
Average hit rate when true positive rate are 50%, 60%, 70%, 80%, 90% and 100%.
a
b
Table 2 summarize the average hit rate of true positive rates 50%, 60%, 70%, 80%, 90% and
100% in dataset 1: sequences with PROSITE pattern, and database 2: sequences with PROSITE
pattern and SWISS-PROT annotations for profile verification. We observed that whether in
dataset 1 or dataset 2, the hit rate of PROSITE patterns are all higher than pattern candidates.
Thus, the PROSITE pattern is really meaningful for protein sequences which have these
PROSITE patterns. However, we also observed that the hit rates in dataset 2 are generally higher
than hit rates in dataset 1. Because dataset 1 only contains sequences with PROSITE patterns but
database 2 contains sequences with PROSITE pattern and SWISS-PROT annotations, it tell us
that the profiles we generated from multiple alignments of ATP-binding proteins may be more
meaningful for protein sequences with “ATP-binding” annotations in SWISS-PROT database.
In Table 3, we summarized the true-positive rates, profile scoring scores, and z-score of profile
scoring scores of top 100, 500, 1000, 1500, 2000, 2500 and 3000 ranked sequences in profile
scoring ranking list. We also compare the hit rates between pattern candidates and PROSITE
patterns. We observed when protein sequences with profile scoring score 0.600, the true positive
rate is 82.27% and the z-score is 2.87. Thus when protein sequences with profile scoring score
higher than 0.600, we can say these protein sequence may be ATP-binding proteins with 82.27%
confidence. When comparing with the hit rate of our defined pattern candidates and PROSITE
patterns, we observed that almost all the top 3000 ranked protein sequences with “ATP-binding”
annotations were all searched by pattern candidates. Although some of pattern candidates
partially overlapped with PROSITE patterns, it tells us that the pattern candidates are useful for
protein function prediction in ATP-binding proteins.
84
Computer Science & Information Technology (CS & IT)
Table 3. Hit rate comparison of pattern candidates and PROSITE patterns in protein function prediction of
ATP-binding proteins
No. of top
ranked
sequence a
True-positive rate
100
500
1000
1500
2000
2500
3000
100.00% (100)
98.40% (492)
95.70% (957)
82.27% (1234)
76.65% (1533)
70.28% (1757)
61.53% (1846)
b
Profile
scoring
score c
0.840
0.720
0.650
0.600
0.583
0.567
0.556
Z-score of
profile
scoring
scored
6.52
4.70
3.63
2.87
2.61
2.37
2.20
Hit rate of all
pattern candidates e
Hit rate of
PROSITE
pattern f
100.00% (100)
100.00% (492)
99.79% (955)
97.65% (1205)
80.43% (1503)
94.25% (1656)
94.53% (1745)
0.00% (0)
0.00% (0)
0.21% (2)
2.35% (29)
19.57% (30)
5.75% (101)
5.47% (101)
The top ranked sequence number. For example, 100 in this column means the 100 ranked sequences with
highest profile scoring score in profile scoring ranking list of ATP-binding protein prediction.
b
True positive rate of protein sequences with “ATP-binding” annotations of top 100, 500, 1000, 1500,
2000, 2500 and 3000 ranked sequences in profile scoring ranking list of ATP-binding protein prediction.
c
The smallest profile scoring scores of top 100, 500, 1000, 1500, 2000, 2500 and 3000 ranked sequences.
d
Z-score of profile scoring scores. The average of all SWISS-PROT sequence scores is 0.411515; the
standard deviation of all SWISS-PROT sequence scores is 0.065701.
e
Hit rate of true positives that predicted by profiles of our defined pattern candidates.
f
Hit rate of true positives that predicted by profiles of PROSITE patterns.
a
3.2 HEM-binding proteins
Heme is a member of a family of compounds called porphyrins, which consist of four pyrrole
rings. Heme metabolism is an important metabolic pathway because many important
hemoproteins contain heme as a prosthetic group. For example, hemoglobin is a very important
hemoprotein and it is an oxygen carrier in the blood. There are also cytochromes, which
participate in important electron transfer reactions, and tryptophan oxygenase which is a
hemoprotein of intermediary metabolism.
3.2.1 Structure similarity matrix and conservation residues
Our results show that structure similarity matrix and SCOP classifications of 40 non-redundant
HEM-binding domains. The structure similarity matrix is still similar with SCOP classification
and MuLiSA can apply to different kinds of ligand-binding proteins. We have also identified
several conservation residues of protein domain clusters in HEM-binding proteins. Figure 4
shows the multiple ligand-bound structure alignment result and identified conservation residues in
“Cytochrome b5 family” of HEM-binding domains. In Figure 4A, the identified conservation
residues are closed to heme in three-dimensional space. It implies that these conservation residues
may play important role in HEM-binding. In Figure 4B, the labeled residue numbers were
belonged to protein domain d1cyo__, which is the selected alignment center C of this cluster, and
the red framed region means the PROSITE patterns. We observed that most identified
conservation residues were on these region, it also tell us that identifying pattern candidates from
conservation residues extension may be a reasonable approach.
Computer Science & Information Technology (CS & IT)
85
A
B
Figure 4. MuLiSA result and identified conservation residues in “Cytochrome b5 family” of HEM-binding
domains. (A) Three-dimensional distributions of identified conservation residues and the ligand
superimposition. Yellow: d1cyo__; blue: d1b5m__; green: d1cxya_; red: d1icca_; grey: d1mj4a_; (B)
Multiple ligand-bound structure alignment result of “Cytochrome b5 family” domains. The identified
conservation residues are closed to heme in three-dimensional space. It implies that these conservation
residues may play important role in HEM-binding. The labeled residue numbers were belonged to protein
domain d1cyo__, which is the selected alignment center C of this cluster, and the red framed region means
the PROSITE patterns.
Table 4. Hit rate comparison of pattern candidates and PROSITE pattern in protein function prediction of
HEM-binding proteins
Top number of
sequence a
True-positive rate b
100
200
300
400
500
600
700
92.00% (92)
80.50% (161)
69.00% (207)
69.75% (279)
70.40% (352)
60.33% (362)
57.86% (405)
Profile
scoring score
c
0.798
0.744
0.708
0.692
0.685
0.685
0.669
Z-score
of
profile scoring
score d
4.72
4.00
3.52
3.30
3.21
3.21
2.99
Hit
rate
of
PROSITE pattern
Hit rate of all
pattern candidates e
f
100.00% (92)
96.27% (155)
97.10% (201)
87.81% (245)
90.34% (318)
90.61% (328)
91.60% (371)
0.00% (0)
3.73% (6)
2.90% (6)
12.19% (34)
9.66% (34)
9.39% (34)
8.40% (34)
The top ranked sequence number.
True positive rate of protein sequence with “Heme” annotations in SWISS-PROT database of top ranked
sequences.
c
The smallest profile scoring scores of top 100, 200, 300, 400, 500, 600 and 700 ranked sequences.
d
Z-score of profile scoring scores. The average of all SWISS-PROT sequence scores is 0.436928; the
standard deviation of all SWISS-PROT sequence scores is 0.071717.
e
Hit rate of true positives that predicted by profiles of our defined pattern candidates.
f
Hit rate of true positives that predicted by profiles of PROSITE patterns
a
b
86
Computer Science & Information Technology (CS & IT)
In order to verify the effectiveness of profiles generated from our alignments in protein function
prediction, we also compare the performance in profile search between datasets 1, which contains
protein sequences with PROSITE pattern; and dataset 2, which contains protein sequences not
only with PROSITE pattern but also have “Heme” annotations in SWISS-PROT database. In
Table 4, we summarized true-positive rates, profile scoring scores, and z-score of profile scoring
scores of top 100, 200, 300, 400, 500, 600 and 700 ranked sequences in profile scoring ranking
list. We also compared with the hit rate of pattern candidates and PROSITE patterns. We
observed that when protein sequences with profile scoring score 0.744, the true positive rate is
80.50% and the z-score is 4.00. Thus when protein sequences with profile scoring score higher
than 0.744, we can say these protein sequences may be HEM-binding proteins with 80.50%
confidence. When comparing the hit rate between pattern candidates and PROSITE patterns, we
observed that almost all the top 700 ranked protein sequences with annotations were searched by
pattern candidates. Although some of pattern candidates may partially overlap with PROSITE
patterns, it tells us the pattern candidates are useful in protein function prediction of HEMbinding proteins.
3.3 Comparison with CE and CLUSTALW
Because multiple ligand-bound structure alignments only focus on ligand-binding sites, we
neglect noise from protein structure apart from the ligand-binding sites and get the functionaldependent alignments of ligand-binding domains. CE and CLUSTALW are structural alignment
and multiple sequence alignment tools, respectively. In Figure 5A, we find that only the
alignments of MuLiSA can align together the PROSITE defined patterns together of two
domains, d1maua_ and d1gtra2. In Figures 5B and 5C, we find that the shift of conservation
patterns of CE alignment result. In fact, for CE uses only protein structure information to undergo
structure alignment, we find that in this case the bad alignment of conservation patterns was
because of a huge structure similar region apart from ATP-binding site, and it did disturb the
alignment of PROSITE patterns. In other words, through ligand superimposition can only focus
on ligand-binding sites and disperse noises from other region, thus the identified conservation
residues and patterns will be much more related to ligand-binding.
A
B
C
Figure 5. The comparison of MuLiSA, CE, and CLUSTALW results of two Class I aminoacyl-tRNA
synthetases (RS), catalytic domains: d1maua_ and d1gtra2. (A) Alignment comparison between three
methods. The shadowed region is the PROSITE defined patterns; (B) 3D structure alignment result of
MuLiSA; (C) 3D structure alignment result of CE. The PROSITE defined patterns together (PROSITE
pattern:
Computer Science & Information Technology (CS & IT)
87
P-x(0,2)-[GSTAN]-[DENQGAPK]-x-[LIVMFP]-[HT]-[LIVMYAC]-G-[HNTG]-[LIVMFYSTAGPC])
of two domains, d1maua_ and d1gtra2. (B), two ATPs were nearly superimposed and the PROSITE
patterns also aligned well. (C) The PROSITE patterns were shifted. In fact, for CE uses only protein
structure information to undergo structure alignment, we find that in this case the bad alignment of
conservation patterns was because of a huge structure similar region apart from ATP-binding site, and it did
disturb the alignment of PROSITE patterns.
4. CONCLUSIONS
We have developed MuLiSA, a multiple ligand-bound structure alignment technique, based on
functional-dependent ligand information to evaluate residue and pattern conservation. The main
difference between our tool and others is that we first superimpose the ligands of proteins but not
protein itself. In this way, the ligand-binding sites are superimposed naturally. Then we could
identify the conservation residues and pattern candidates according to these positions and
segments which were superimposed along with ligands. We have applied MuLiSA to ATPbinding proteins and HEM-binding proteins. MuLiSA can identify conservation residues and
pattern candidates which play important role in ligand-binding or binding site conformation
stability. For predicting protein conserved residues and patterns, our MuLiSA achieved high
accuracies 80.5% and 82.3 for HEM-binding and ATP-binding proteins, respectively. We believe
that our MuLiSA is very useful for prediction of conserved results, functional motifs, and protein
functions.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
Page, R. D. M., And Holmes, E. C. (1998) Molecular Evolution: A Phylogenetic Approach, Pp 228279, Blackwell, Oxford.
Falquet, L., Pagni, M., Bucher, P., Hulo, N., Sigrist, C. J., Hofmann, K., And Bairoch, A. (2002) The
Prosite Database, Its Status In 2002, Nucleic Acids Research 30, 235-238.
Bateman, A., Coin, L., Durbin, R., Finn, R., Hollich, V., Griffiths-Jones, S., Khanna, A., Marshall,
M., Moxon, S., Sonnhammer, E. L., Studholme, D. J., Yeats, C., And Eddy, S. R. (2004) The Pfam
Protein Families Database, Nucleic Acids Research 32, D138-D141.
Casari, G., Sander, C., And Valencia, A. (1995) A Method To Predict Functional Residues In
Proteins, Nature Structural Biology 2, 171-178.
Pietrokovski, S., Henikoff, J. G., And Henikoff, S. (1996) The Blocks Database--A System For
Protein Classification, Nucleic Acids Research 24, 197-200.
Jones, S., And Thornton, J. M. (1997) Prediction Of Protein-Protein Interaction Sites Using Patch
Analysis, Journal Of Molecular Biology 272, 133-143.
Shatsky, M., Nussinov, R., And Wolfson, H. J. (2002) Flexible Protein Alignment And Hinge
Detection, Proteins: Structure Function And Genetics 48, 242-256.
hindyalov, I. N., And Bourne, P. E. (1998) Protein Structure Alignment By Incremental
Combinatorial Extension (Ce) Of The Optimal Path, Protein Engineering 11, 739-747.
Higgins, D., Thompson, J., Gibson, T., Thompson, J. D., Higgins, D. G., And Gibson, T. J. (1994)
Clustalw: Improving The Sensitivity Of Progressive Multiple Sequence Alignment Through Sequence
Weighting,Position-Specific Gap Penalties And Weight Matrix Choice, Nucleic Acids Research 22,
4673-4680.
Sali, A., Overington, J. P., Johnson, M. S., And Blundell, T. L. (1990) From Comparisons Of Protein
Sequences And Structures To Protein Modelling And Design, Trends In Biochemical Sciences 15,
235-240.
Lichtarge, O., Bourne, H. R., And Cohen, F. E. (1996) An Evolutionary Trace Method Defines
Binding Surfaces Common To Protein Families, Journal Of Molecular Biology 257, 342-358.
Innis, C. A., Shi, J., And Blundell, T. L. (2000) Evolutionary Trace Analysis Of Tgf-Beta And
Related Growth Factors: Implications For Site-Directed Mutagenesis, Protein Engineering 13, 839847.
Landgraf, R., Xenarios, I., And Eisenberg, D. (2001) Three-Dimensional Cluster Analysis Identifies
Interfaces And Functional Residue Clusters In Proteins, Journal Of Molecular Biology 307, 14871502.
88
Computer Science & Information Technology (CS & IT)
[14] Notredame, C., Higgins, D. G., And Heringa, J. (2000) T-Coffee: A Novel Method For Fast And
Accurate Multiple Sequence Alignment., Journal Of Molecular Biology 302, 205-217.
[15] Altschul, S. F., Gish, W., Miller, W., Myers, E. W., And Lipman, D. J. (1990) Basic Local Alignment
Search Tool, Journal Of Molecular Biology 215, 403-410.
[16] Holm, L., And Sander, C. (1993) Protein Structure Comparison By Alignment Of Distance Matrices,
Journal Of Molecular Biology 233, 123-138.
[17] Gibrat, J. F., Madej, T., And Bryant, S. H. (1996) Surprising Similarities In Structure Comparison,
Current Opinion In Structural Biology 6, 377-385.
[18] Laskowski, R. A., Hutchinson, E. G., Michie, A. D., Wallace, A. C., Jones, M. L., And Thornton, J.
M. (1997) Pdbsum: A Web-Based Database Of Summaries And Analyses Of All Pdb Structures,
Trends In Biochemical Sciences 22, 488-490.
[19] Murzin, A. G., Brenner, S. E., Hubbard, T., And Chothia, C. (1995) Scop: A Structural Classification
Of Proteins Database For The Investigation Of Sequences And Structures, Journal Of Molecular
Biology 247, 536-540.
[20] Boeckmann, B., Bairoch, A., Apweiler, R., Blatter, M. C., Estreicher, A., Gasteiger, E., Martin, M. J.,
Michoud, K., O'donovan, C., Phan, I., Pilbout, S., And Schneider, M. (2003) The Swiss-Prot Protein
Knowledgebase And Its Supplement Trembl In 2003, Nucleic Acids Research 31, 365-370.
[21] Berman, H. M., Westbrook, J., Feng, Z., Gilliland, G., Bhat, T. N., Weissig, H., Shindyalov, I. N.,
And Bourne, P. E. (2000) The Protein Data Bank, Nucleic Acids Research 28, 235-242.
[22] Besl, P. J., And Mckay, N. D. (1992) A Method For Registration Of 3-D Shapes, Ieee Transactions
On Pattern Analysis And Machine Intelligence 14, 239-256.